STANDARD PROTOCOL #1 FOR AUTOMATED CVR
Submitted to
REGIONAL TECHNICAL FORUM
Submitted by
PCS UtiliData
6620 N. Market St.
Spokane, WA 99217
November 1, 2011
Prepared by PCS UtiliData
STANDARD PROTOCOL #1 FOR AUTOMATED CVR
TABLE OF CONTENTS
1.
PURPOSE ................................................................................................... 1
2.
SUNSET CRITERIA ......................................................................................... 1
3.
DEFINITION
4.
ELIGIBLE PROJECTS ....................................................................................... 2
5.
REQUIRED KNOWLEDGE AND SKILLS OF PRACTITIONER .......................................... 2
6.
DATA COLLECTION REQUIREMENTS .................................................................. 2
a.
b.
c.
d.
OF KEY TERMS ............................................................................ 1
Temperature .................................................................................................................... 2
Metering and Data ........................................................................................................... 3
Instrumentation ............................................................................................................... 3
Shop Calibration and Field Verification ........................................................................... 3
7.
POST-PERIOD RE-VERIFICATION TRIGGERS ......................................................... 3
8.
BASELINE PERFORMANCE ............................................................................... 3
a.
9.
Application Specifications: .................................................Error! Bookmark not defined.
METHOD DESCRIPTION: ................................................................................ 4
a.
b.
c.
d.
Primary Method ............................................................................................................... 4
Program Savings............................................................................................................... 4
CVRf Estimation (applies to the performance evaluation period) ................................... 5
Automated CVR Performance Forecasting ...................................................................... 5
10.
MODEL ..................................................................................................... 5
11.
CONTROL GROUP: ....................................................................................... 6
12.
RELATIONSHIP TO OTHER PROTOCOLS AND GUIDELINES ......................................... 6
13.
RECOMMENDED MODELS AND TOOLS: .............................................................. 6
14.
REFERENCES: .............................................................................................. 6
i
STANDARD PROTOCOL #1 FOR AUTOMATED CVR
1. PURPOSE
This protocol establishes a method by which annual electrical energy savings (kWh) can be estimated for
electric utility distribution feeders and substations feeding residential, commercial and industrial
customers when those feeders or substations are equipped with an automated CVR system.
This protocol specifies minimum acceptable data collection requirements and the method by which
these data are to be used in computing savings.
2. SUNSET CRITERIA
This protocol is approved for use until December 31, 2020 at which time it shall be reviewed for
continued approval by the RTF.
3. DEFINITION OF KEY TERMS
Automated CVR System: Automated CVR (ACVR) systems are also known as Volt/VAR Optimization
(VVO) systems, Integrated Volt/VAR Control (IVVC) systems and by other names and acronyms when
they are being operated in a CVR mode. They key feature is that they be considered a “smart grid”
Distribution Automation (DA) application.
Conservation Voltage Regulation/Reduction (CVR): CVR is the operating practice of controlling
distribution feeder voltage so that utilization devices operate at their peak efficiency, which is usually at
a level near the lower bounds of their utilization and/or nameplate voltages.
Conservation Voltage Regulation/Reduction Factor (CVRf): CVRf = per unit (or %) energy saved per unit
(or %) voltage reduced. The CVRf represents the average slope on the utilization device’s efficiency
curve between the current voltage and the new proposed regulated voltage. For example, a CVRf of
1.00 would essentially indicate a 1% reduction in energy usage for every 1% reduction in voltage.
Project. An automated CVR system is installed on a utility’s distribution feeder or feeders feeding
residential, commercial and industrial customers.
Baseline. Each feeder acts as its own baseline in the protocol on ACVR off days during the protocol
testing period.
Post Period Re-Verification. This refers to re-verification of CVRfs when certain changes have occurred
to either the feeder load or the physical topology of a feeder.
Commissioning. This is the process of testing and adjustment required to ensure that the automated
CVR system is functioning properly.
1
STANDARD PROTOCOL #1 FOR AUTOMATED CVR
4. ELIGIBLE PROJECTS
Eligible projects include all installations or implementations of automatic CVR systems on utility
substations or feeders where these automated systems can be turned on and off on a daily basis, have
the voltage set-points changed on a daily basis, have the ability to measure and record period average
bus and end-of-line voltages, period kWh, period kVARh on a per feeder basis and measure and record
period average temperature. The method is applicable where no previous energy usage information is
available
The ideal application would be where the automatic CVR control components could also monitor and
store the period data.
5. REQUIRED KNOWLEDGE AND SKILLS OF PRACTITIONER
The practitioner who has lead responsibility for applying this protocol to an automated CVR system must
have at a full understanding of the following:

Appropriate knowledge of the application of CVR to distribution systems and the underlying physics
of relationships between operating voltage levels and energy consumption

Appropriate knowledge of the use of engineering time series analysis

Appropriate knowledge of the use of Robust Statistical procedures used to analyze non-Gausian
data

Appropriate knowledge of distribution feeder and substation operations

Understand the requirements and procedures of this protocol
The practitioner must also be able to successfully:

Operate in a MatLab® M-Code environment

Inspect and interpret raw feeder energy and voltage data
6. DATA COLLECTION REQUIREMENTS
a.
Temperature
Correct temperature data is essential to the accurate use of this verification method. It is
recommended that the automated CVR system records period temperatures at the substation.
Because the substation is usually at the geographic center of the area served this temperature will
usually suffice. However, if significant microclimates are known to exist, temperature monitoring
and recording may also be required at the feeder end-of-line location, so that an average
temperature for the feeder may be obtained.
2
STANDARD PROTOCOL #1 FOR AUTOMATED CVR
b.
Metering and Data
Data recording periods should be no greater than one hour, and can be as short as the system
allows. Weather data should be collected on the same time period as the load data. Data collected
is subject to audit.
c.
Instrumentation
Voltage monitors should have linearity of better than ½% within the expected ranges of voltage and
temperature drift should be less than ½% from -40 degrees C to 65 degrees C. Power monitors
should be revenue grade accuracy but need not be revenue class.
d.
Shop Calibration and Field Verification
Instruments and meters should be shop calibrated. Field verification and inspections are required to
verify correct installation and correct readings.
7. POST-PERIOD RE-VERIFICATION TRIGGERS
Re-verification will be required when there is a +/- 10% shift in temperature adjusted total annual load
or a +/-10% shift in temperature adjusted total load during heating regime hours or a +/-10% shift in
temperature adjusted total load during cooling regime hours or permanent reconfiguring of the
distribution system (not including re-conductoring, transformer replacement, capacitor banks, or other
distribution system efficiency project.)
If re-verification is triggered by a shift in the loads during heating or cooling regimes, the re-verification
protocol will consist of one sixty day period during either the heating or cooling period. On alternate
days the system is at full voltage reduction, and the next day at the controlled nominal midpoint.
If re-verification is triggered by a shift in total annual load or a permanent re-configuration the reverification protocol will consist of two sixty day periods, one in the heating period and one in the
cooling period. On alternate days the system is at full voltage reduction, and the next day at the
controlled nominal midpoint.
The new CVRfs determined by these re-verifications will be used in lieu of the original CVRfs.
8. BASELINE PERFORMANCE
The baseline voltage levels are established by the historical regulator or LTC control settings. One or
more years of historical regulator or LTC setting information should be made part of the verification data
records.
3
STANDARD PROTOCOL #1 FOR AUTOMATED CVR
9. METHOD DESCRIPTION:
a.
Primary Method
The primary method of verifying energy savings is to operate the system in such a way as to operate
at different voltage levels on alternating days. The initial verification period would last one year. The
verification would begin with 90 days or three months of one day at full voltage reduction and one
day of automated CVR off. During the next 9 months the automated CVR would be on continuously
except for three (non-contiguous) of these nine months selected based on season and other factors
such as geographic weather patterns etc., to operate the system so that on alternate days the
system is at full voltage reduction, and the next day with automated CVR off.
Time series analysis procedures, robust statistical methods, and temperature compensation
methods are then used to evaluate the total energy conservation by comparing energy use on
similar days at different voltage levels. For instance, winter weekdays would be compared against
winter weekdays, summer weekdays against summer weekdays, etc. Conservation voltage
regulation factors (CVRf) then computed for each feeder the different seasons for weekdays and for
weekends.
CVRfs are used to estimate total ongoing energy conservation. CVRfs are verified during similar
periods in following periods by running alternating days with full end of line voltage reduction and 2
volts above full end of line voltage reduction for two to four week periods.
b.
Program Savings
The program savings are estimated by using the following definition:
Esaved = Eused [(CVRf*Vr%/(1-CVRf*Vr%)]
In which:
Esaved = Energy Conserved for period in kWh, MWh or GWh
Eused = Measured Energy used for period in kWh, MWh or GWhCVRf = Period conservation
voltage reduction factor as computed using time series analysis and robust statistical methods
with temperature compensation for specific seasons. CVRf will be different for weekday and
weekend. (See estimation method below.)
Vr = Average period end of line voltage reduction
Vr% = Average period end of line voltage reduction in percent
Voc = measured average end of line voltage with automated CVR non operational
Vcvr = measured average end of line voltage with automated CVR operational
4
STANDARD PROTOCOL #1 FOR AUTOMATED CVR
Vr = Voc – Vcvr
Vr% = Vr/Voc * 100
c. CVRf Estimation (applies to the performance evaluation
period)
Integrated demand profiles, one each for the automated CVR system active and inactive, are
estimated on a common ambient temperature basis using the method (“Estimation of Automated
CVR System Performance Using Observed Energy Demand Load Profiles”); the 24-hour sum of the
difference between these profiles is the estimated conserved energy for the evaluation. The mean
difference of the end of circuit voltages for the automated CVR system active and inactive is
estimated. The CVRf is then determined from the ratio of these two quantities, and can be
expressed on an absolute or per unit basis (the per unit basis is recommended).
Recognizing (1) the stochastic nature of the energy observations as discussed in the UtiliData CVR
Estimation Method, (2) the requirement to evaluate the performance of candidate automated CVR
systems using the smallest (least duration) set of energy observations, and (3) that the probability
densities of the relevant observations clearly exhibit non-homogeneous variance and are also clearly
not Gaussian processes, the required estimations should be carried out using robust statistical
procedures. Specifically, the Minimum Covariance Determinant estimators should be applied,
because (1) their breakdown point is high and (2) they do not require that the observations exhibit a
symmetrical probability density.
d.
Automated CVR Performance Forecasting
The UtiliData CVR Estimation Method referenced above, estimates CVR using the observations of
the automated CVR system inactive state as a reference. In principle, forecasting for a given circuit
then simply requires a base demand profile, a projected end of circuit voltage reduction, and the
estimation results from the evaluation period.
10. MODEL
The current model used for the time series analysis includes compensation for temperature. There are a
number of additional factors that affect energy use and could be added to the model. Addition of these
factors will tend to improve the predictive accuracy and reduce “outlier” data points. Factors that may
be considered for inclusion in the model in the future will include daylight and dark hours, solar
intensity, day of week, humidity, etc.
Adding any or all of these to the model should not change the basic measurement and verification
protocol.
5
STANDARD PROTOCOL #1 FOR AUTOMATED CVR
11. CONTROL GROUP:
No control group or baseline is required because with on-off and variable voltage set point capability,
the application group can act as its own control group or baseline during testing periods.
12. RELATIONSHIP TO OTHER PROTOCOLS AND GUIDELINES
The relationship between this protocol and other relevant protocols and guidelines is as follows:

International Performance Measurement and Verification Protocol, Concepts and Options for
Determining Energy and Water Saving, Volume 1, Revised March 2002. DOE/GO-102002-1554. This
protocol is consistent with Options C - and D of that document.

M&V Guidelines: Measurement and Verification for Federal Energy Projects Version 3.0, U.S.
Department of Energy Federal Energy Management Program. This protocol is consistent with
Option C.
13. RECOMMENDED MODELS AND TOOLS:
UtiliData® Automated CVR Estimation Method Tools
MatLab® (©1994-2003 by the MathWorks, Inc.) tools are available from PCS UtiliData to use with this
protocol.
14. REFERENCES:
Rousseeuw, P J, Leroy AM, ‘Robust Regression and Outlier Detection’, Wiley 1987.
Rousseeuw, P J, 'Introduction to Positive Breakdown Methods', in Handbook of Statistics, Volume 15:
Robust Inference, editors G S Maddala and C R Rao, Elsevier 1997.
“Estimation of Automated CVR System Performance using Observed Energy Demand Profiles”, David
Bell, March 15, 2004. (available at
http://www.pcsutilidata.com/userfiles/file/CVR_Performance_Estimation_2004.pdf)
6